No.

Consist deviated under the two circumstances is pretty well studied. A substantial number of readings in a doctors office are high. Within certain margins it's accounted for, obviously there are people that exhibit the effect more than others.

But taking blood pressure correctly is generally not difficult, can be readily trained for and if necessary can be demonstrated easily by both parties.

Getting results which are consistent, wrong, but approximately in the correct range is actually quite hard.

ML has always been a category of AI.

I've got books from decades ago referencing ML as The AI category's biggest wins.

Also, using AI to train a model for categorization of data has been with us since the 80s. Its effectiveness has increased of course.

Terms of service aren't much of a concern here. Honestly, completely irrelevant unless OpenAI opens up 30,000 streams at once.

Terms of service have little value in court when it comes to scraping of content that doesn't cause issues with the service itself. Contract violation with near zero repercussions.

This type of thing is actually very hard to test without introducing your own bias.

It's fun watching people when they introduce their own bias and then claim it's the system.

The level found is largely irrelevant.

You might not want it, and as policy it's great to get them out of the supply chain; but at these concentrations a lifetime of heavy drinking will increase your risk of dying by bicycle (as a pedestrian) higher than the increase risk from this source.

Beer isn't distilled.

Based on the company selling them as if they knew exactly what they were; I'm wondering if they actively participated in the process and "accidentally" took them with other stuff that was intended to go.

This decision looks like a precursor to outright corruption.

Have you ever looked at the specs?

Not something I'd claim ownership in, even if I had contributed.

Destroying your competition is a time honour tradition with companies.

Occasionally someone will restart a similar company, but usually the non-competes the original founders have when bought out are sufficient to prevent that. (Money and time change people's priorities, so having a 2 year industry reaction ends up being a lifetime of change).

Of course companies gut the assets then can use elsewhere, retailing the value they can; but destroying companies the feel compete within their markets is usually the primary goal.

When a company disposes of another company at the 5yr mark it usually means that was the plan from day 1. Buying up companies and dismantling them right away can get your company scrutinized, but after 5 years "it just didn't work out" is an accepted answer to destroying valuable assets. (Especially when they can spend the time devaluing them too)

So you think the audio book industry loves these types of podcasts? Not at all.

Keep digging on the blue / green and you'll find the same general issues apply across cultures based on historical events.

Japan didn't have a reason to prioritize differentiation, they perceive the color the same as you. It's the translation from stimulus to specific categorization that is different.

Xkcd color chart comparison between male / female color names highlights something similar. It's not perception though, it's "importance". I don't care if there's 4 shades of green by name, or 72. Someone else with a keen interest could choose to differentiate in 256 unique names. It doesn't however mean we perceive the color differently.

Unsurprised. I've had this conversation multiple times with different people, not surprising since my day job is web development and my hobbies include photography.

With color blindness excepted and probably slightly interesting to compare between, perceptual evidence provides us with color relationships that are obviously very similar in most instances.

There's arguments over color names, but if you look deeper you'll find a long list of consistencies (check out the color blue across diverse cultures; the similarities are fascinating).

(Color nameing groups seem to be structured importance of differentiation; those priorities are different between genders, but provide a perceptual color chart and ask people to sort them and they'll generally come up the same.

I have a site with thousands of products and massive search space due to the number of attributes customers can search on.

The bots are completely stupid trying to product search and iterate over criteria combinations which aren't overall very useful.

You can efficiently grab all our products if a small amount of intelligence is applied and you'd have negligible impact on the server. Instead the bots try to search using each and every possible attribute and therefor blow the cache. I could write a bot to fetch the entire contents of the site in less than an hour and the server wouldn't take a hit; the regular search bots aren't usually pathological about the search criteria.

If the AI ones paid attention to the robots file, and site map, they'd barely be noticed either.

Sure, I can restrict the methods used to search the site, limit it to a cachable subset of search terms, but that would negatively impact users that have unusual requirements.

It's interesting because creating fine literature doesn't pay up front.

You'll spend years fine tuning your work, writing it, rewriting it while making no money from it. Why not work on fixing other peoples mess in the meantime and rake in the dough.